Construct a Standardized ROI Result — roi

Creates a uniform return type for per-ROI analysis. Every fit_roi method should return an roi_result.

Usage

roi_result(
  metrics,
  indices,
  id,
  result = NULL,
  error = FALSE,
  error_message = "~",
  warning = FALSE,
  warning_message = "~"
)

Arguments

metrics: Named numeric vector of performance metrics.
indices: Integer vector of voxel/vertex indices for this ROI.
id: Scalar ROI identifier (center voxel ID or region number).
result: Optional detailed result object (e.g., classification_result).
error: Logical; TRUE if this ROI failed.
error_message: Character error description, or "~" if no error.
warning: Logical; TRUE if a warning was raised.
warning_message: Character warning description, or "~" if none.

Value

A list of class "roi_result".

Examples

# Successful ROI result
res <- roi_result(
  metrics = c(accuracy = 0.85, AUC = 0.9),
  indices = 1:10,
  id = 42
)
res$metrics
#> accuracy      AUC 
#>     0.85     0.90 

# Error ROI result
err <- roi_result(
  metrics = NULL,
  indices = 1:10,
  id = 42,
  error = TRUE,
  error_message = "Too few features"
)
err$error
#> [1] TRUE